Refining Obesity Risk Models: Gradient Boosting for Superior Predictive Performance

It is important to recognize the risk of obesity. In this study, we compared three machine learning models to see how well they can detect the risk of obesity. The best model was the XGBoost classifier, which was 91% accurate. The other two models, LightGBM and CatBoost, were a bit less accurate, each at $\mathbf{9 0 \%}$. These models could be important for identifying people at risk of obesity. They can help create personalized healthcare plans to prevent obesity and its serious related health issues. This helps them find small patterns that could show a risk of obesity. This information can help people create personalized ways to prevent and manage obesity. Finding the right people who are at risk of becoming overweight could be an important finding to improve public health. Person-time detection can help find people who are at risk so that we can focus on preventing problems.